https://doi.org/10.1177/01461672211025945

Personality and SocialPsychology Bulletin 1 –19© 2021 by the Society for Personalityand Social Psychology, IncArticle reuse guidelines: sagepub.com/journals-permissionsDOI: 10.1177/01461672211025945journals.sagepub.com/home/pspb

Empirical Research Paper

Feeling and expressing gratitude are integral to daily life and result in greater daily positive affect and self-esteem (Kashdan et al., 2006), increased job satisfaction (Waters, 2012), and higher levels of various indicators of well-being (Emmons & McCullough, 2003). Furthermore, feeling and reflecting on gratitude have been shown to increase positive affective well-being (Kaplan et al., 2014) and decrease per-ceived stress and depressive symptoms (Cheng et al., 2015). In organizations, leaders’ gratitude expression can spark a ripple of reciprocity as employees experience increased positive emotions (Algoe et al., 2016) and feelings of social worth (Cho & Fast, 2012).

Although researchers have documented numerous down-stream consequences of feeling and expressing gratitude, less is known about the antecedents of these important out-comes. Existing work has tended to focus on the nature of the helping task (Lee et al., 2018) while overlooking the inter-personal relationship between the favor giver and the favor recipient and the intrapersonal psychology of the favor recip-ient. This is a problematic shortcoming because scholars have pointed out that “gratitude is largely influenced by the social context where helping exchanges take place and not all instances of help necessarily lead to receipt of gratitude” (Lee et al., 2018, p. 3; see also Algoe et al., 2010). Thus, an open and critical question is why some instances of help

result in the recipient feeling and expressing gratitude while others do not.

We propose that the power dynamic between two indi-viduals is a crucial predictor of feeling and expressing grati-tude. According to the social distance theory of power (Magee & Smith, 2013), the asymmetric dependence that characterizes power relations produces asymmetric social distance, with relatively lower-power (vs. higher-power) individuals demonstrating more concern with and respon-siveness to the behaviors and characteristics of their inter-action partners. In addition, this asymmetric dependence can generate feelings of entitlement because high-power indi-viduals expect low-power individuals to appease them in various ways (Magee & Smith, 2013).

We elaborate on the social distance theory of power by examining two mechanisms linking relative power to expres-sions and feelings of gratitude. Specifically, we propose that

1025945 PSPXXX10.1177/01461672211025945Personality and Social Psychology BulletinAnicich et al.research-article2021

1University of Southern California, Los Angeles, USA2Cornell University, Ithaca, NY, USA3Columbia University, New York City, NY, USA

Corresponding Author:Eric M. Anicich, Marshall School of Business, University of Southern California, 3670 Trousdale Pkwy, Los Angeles, CA 90089, USA. Email: anicich@marshall.usc.edu

Thanks, but No Thanks: Unpacking the Relationship Between Relative Power and Gratitude

Eric M. Anicich1 , Alice J. Lee2, and Shi Liu3

AbstractPower and gratitude are universal features of social life and impact a wide range of intra- and interpersonal outcomes. Drawing on the social distance theory of power, we report four studies that examine how relative power influences feelings and expressions of gratitude. An archival analysis of author acknowledgements in published academic articles (N = 1,272) revealed that low-power authors expressed more gratitude than high-power authors. A pre-registered experiment (N = 283) involving live conversations online found that having relatively low power caused increased feelings and expressions of gratitude after benefiting from a favor. Another pre-registered experiment (N = 356) demonstrated that increased interpersonal orientation among lower power individuals and increased psychological entitlement among higher power individuals drove these effects. Finally, an archival analysis of conversational exchanges (N = 136,215) among Wikipedia editors revealed that relational history moderated the effect of relative power on gratitude expression. Overall, our findings highlight when and why relative power influences feelings and expressions of gratitude.

Keywordspower, gratitude, entitlement, interpersonal orientation, hierarchy

Received January 18, 2020; revision accepted May 18, 2021

2 Personality and Social Psychology Bulletin 00(0)

lower-power favor recipients experience a greater interper-sonal orientation than higher-power favor recipients which increases their expression and feeling of gratitude, whereas higher-power favor recipients experience a greater sense of psychological entitlement than lower-power favor recipients which decreases their expression and feeling of gratitude. In addition, we embrace perspectives that highlight the inher-ently interpersonal nature of power (Magee & Smith, 2013; Smith & Magee, 2015) in testing a moderator of the effect of relative power on gratitude expression: the relational history between the two parties.

Relative Power, Gratitude Expression, and Felt Gratitude

Central to our theorizing are theoretical perspectives and empirical insights suggesting that low power focuses indi-viduals outward toward others, whereas high power increases an inward focus on the self and one’s own contributions (Galinsky et al., 2015; Magee & Smith, 2013; Rucker et al., 2018; Williams, 2014). These relational perspectives on power align with the inherently relational nature of gratitude which, according to the find-remind-and-bind theory of grat-itude (Algoe, 2012), “helps one to find new or remind one of current good relationship partners and to bind the two more strongly together” (Algoe et al., 2016, p. 1). Importantly, the find-remind-and-bind theory of gratitude emphasizes a rela-tional appraisal of the situation, which departs from prior economic accounts of gratitude (Algoe, 2012). We propose that power—reflected in the degree of interdependence between interaction partners (Kelley et al., 1983; Kelley & Thibaut, 1978; Magee & Smith, 2013)—is an important component of that relational appraisal and is likely to influ-ence both expressions and feelings of gratitude.

Gratitude expression serves instrumental purposes for low-power individuals due to the asymmetric dependence that characterizes power relations (Emerson, 1962; Fiske & Berdahl, 2007; Magee & Galinsky, 2008). To meet their goals, low-power individuals depend crucially on the percep-tions and evaluations of higher-power others. As a result, low-power individuals are likely to be especially attuned to the behaviors and expectations of higher-power others who have the authority to punish and reward them (Keltner et al., 2003). Thus, lower-power favor recipients are likely to be especially motivated to express large amounts of gratitude to higher-power favor givers as a means of meeting normative expectations and avoiding punishment. Therefore, we pro-pose the following hypothesis:

Hypothesis 1: Relative power will be negatively associ-ated with gratitude expression following the receipt of a favor.

Relative power is also likely to influence feelings of gratitude. Indeed, Magee and Smith (2013) propose that

high-power individuals are less likely to experience socially engaging emotions (e.g., gratitude) due to their reduced motivation for affiliation with others. The authors base this claim on an abundance of work showing that high power is associated with less interest in and concern with the mental states of others and in social affiliation more generally (Fiske, 1993; Galinsky et al., 2006; Inesi et al., 2012; Lammers et al., 2012). A complementary view expressed by MacKenzie and Baumeister (2019) holds that feelings of gratitude are driven by an individual’s perception that they received a greater benefit than they had a right to expect. As we describe in greater detail below when developing our mediating Hypothesis 4b, higher-power individuals are more likely than lower-power individuals to expect to receive vari-ous benefits from others and thus may be less likely to feel grateful following the receipt of any particular benefit. Taken together, we propose that experiencing relatively high com-pared to low power will reduce the amount of gratitude favor recipients feel.

Hypothesis 2: Relative power will be negatively associ-ated with feelings of gratitude following the receipt of a favor.

In the next section, we propose two mechanisms linking relative power to expressions and feelings of gratitude.

Interpersonal Orientation and Psychological Entitlement as Simultaneous Mediators

Relative power is likely to influence expressions and feel-ings of gratitude through differences in interpersonal orienta-tion. We focus on interpersonal orientation as a potential mediator given the inherently interpersonal and relational nature of both power (Magee & Smith, 2013; Smith & Magee, 2015) and gratitude (Algoe, 2012). For high inter-personal orientation individuals, “the distribution of power and dependence in the relationship” is particularly relevant (Swap & Rubin, 1983, p. 209). In this way, our conceptual-ization of interpersonal orientation is closely related to Yang et al.’s (2008) conceptualization of relational orientation which reflects “those tendencies that an individual appreci-ates in developing, maintaining, and using interpersonal rela-tionships” (p. 714). Our conceptualization of interpersonal orientation is also closely related to Cheung et al.’s (2001) conceptualization of interpersonal relatedness which reflects “a strong orientation toward instrumental relationships; emphasis on occupying one’s proper place and engaging in appropriate action; avoidance of internal, external, and inter-personal conflict; and adherence to norms and traditions” (p. 425), characteristics which the authors note “are akin to Singelis’s (1994) concept of the interdependent self- construal.” Finally, our conceptualization of interpersonal orientation aligns nicely with Hashimoto and Yamagishi’s

Anicich et al. 3

(2013, p. 143) view of the “interdependent self” as both a “harmony seeker” and a “rejection avoider.” We adapted items from several of these closely related constructs to con-struct our measure of interpersonal orientation in Study 3.

Importantly, past work suggests that lower-power indi-viduals compared to higher power individuals are more interpersonally oriented. For example, less powerful versus more powerful individuals are more receptive to others’ social needs (Galinsky et al., 2006), more interpersonally attentive (Fiske, 1993; Goodwin et al., 2000), and more com-passionate (Van Kleef et al., 2008). Lower-power individuals compared to higher-power individuals also prefer and expe-rience less social distance (Case et al., 2015; Lammers et al., 2012; Magee & Smith, 2013) and have more of an inter-dependent compared to independent self-construal (Lee & Tiedens, 2001; Magee et al., 2010; Smith & Trope, 2006). Related work stipulates that low-power individuals are more likely to demonstrate a communal orientation than high-power individuals (for reviews see, Keltner et al., 2003; Rucker & Galinsky, 2016; Rucker et al., 2018).

On the basis of this work, we propose that elevated inter-personal orientation among low-power individuals is likely to drive gratitude expression because low-power individuals will want to behaviorally signal their interpersonal orienta-tion to the favor giver, thereby advancing the low-power individual’s goal of becoming more closely bound to the favor giver (Algoe, 2012). Furthermore, we propose that elevated interpersonal orientation among low-power indi-viduals is also likely to increase feelings of gratitude, which are likely to emerge following the receipt of a favor from a higher power individual. In summary, elevated interpersonal orientation should, in turn, lead to increased expressions and feelings of gratitude following the receipt of a favor. Thus, we predict:

Hypothesis 3a: Lower-power individuals will express more gratitude than higher-power individuals due to experiencing increased interpersonal orientation.Hypothesis 3b: Lower-power individuals will feel more gratitude than higher-power individuals due to experienc-ing increased interpersonal orientation.

As previously noted, high power increases a sense of self-focus and decreases attentiveness to and concern for inter-personal relationships (Magee & Smith, 2013). Consistent with this line of reasoning, past work has demonstrated that having various forms of power is associated with an increased sense of entitlement (Campbell et al., 2004; Côté et al., 2020; Piff, 2014; Sawaoka et al., 2015). For example, high-power individuals are more likely than low-power individuals to claim a disproportionate share of resources due to an ele-vated sense of entitlement (De Cremer & Van Dijk, 2005). Conversely, the inherent vulnerability associated with lack-ing power is likely to produce a reduced sense of entitlement. In addition, activation of the behavioral approach system (BAS)—which is the neural substrate of high power and

reward-seeking behavior more generally (Keltner et al., 2003)—plays a crucial role in other forms of egocentrism, including narcissism (Miles et al., 2019; Neria et al., 2016), which is directly linked to a sense of entitlement (Raskin & Hall, 1979).

Entitlement may cause favor recipients to feel indifferent toward and/or have difficulty recognizing the contributions that favor givers have made to their own outcomes. As MacKenzie and Baumeister (2019, p., 415) note, “To the extent that one is entitled to some positive outcome, one does not have to be grateful for it.” More specifically, enti-tled people tend to construe the receipt of a benefit as a reward for their own effort instead of another’s benevolent act (Pelser et al., 2015). Indeed, self-focus is likely to con-tribute to the view that one is the sole arbiter of one’s out-comes. Simply put, “gratitude is the opposite of entitlement” (Twenge & Campbell, 2009, p. 240).

We propose that elevated psychological entitlement is likely to reduce gratitude expression because entitlement prompts a desire to behaviorally signal independence from the favor giver to maintain the illusion that their outcomes were self-made. Furthermore, we propose that elevated psy-chological entitlement is also likely to decrease feelings of gratitude because entitled individuals are unlikely to accu-rately perceive the contributions of others, which may pre-vent feelings of gratitude from emerging. Integrating these perspectives, we propose the following:

Hypothesis 4a: Higher-power favor recipients will express less gratitude than lower-power favor recipients due to an increased sense of psychological entitlement.Hypothesis 4b: Higher-power favor recipients will feel less gratitude than lower-power favor recipients due to an increased sense of psychological entitlement.

The Moderating Role of Relational History

We have argued that lower-power individuals compared to higher-power individuals are more likely to be discerning of different aspects of their relationships with others due to both an elevated interest in affiliating with others and a recogni-tion of their greater exposure to interpersonal threats at the hands of higher-power favor givers.

Building on these ideas, we further propose that when interacting with higher-power favor givers, lower-power favor recipients will be more motivated by a punishment avoidance goal than higher-power favor recipients will be when interacting with lower-power favor givers. We base this prediction on the claim that when determining the appro-priate amount of gratitude to express to a higher-power favor giver, overestimating the appropriate amount of gratitude to express is likely to result in less punishment for lower-power favor recipients than underestimating the appropriate amount of gratitude to express. For example, at some point in their relational history with a particular higher-power favor giver,

4 Personality and Social Psychology Bulletin 00(0)

a lower-power favor recipient may initially conclude that it is interpersonally safe enough to express less gratitude to the higher-power favor giver. However, the lower-power favor recipient may nonetheless conclude (perhaps wisely) that the interpersonal risk associated with upsetting their higher- power interaction partner is too great to justify adjusting their level of gratitude expression downward. That is, they may decide to continue to express a potentially unnecessarily large amount of gratitude to the higher-power favor giver because they recognize that the costs associated with overes-timating the appropriate amount of gratitude are relatively trivial compared to the costs associated with underestimating the appropriate amount of gratitude.

This view is consistent with Error Management Theory (Haselton & Buss, 2000) which proposes that the costs asso-ciated with committing Type I errors (e.g., expressing grati-tude when the favor giver does not expect to receive gratitude) and Type II errors (e.g., not expressing gratitude when the favor giver expects to receive gratitude) are often asymmetri-cal. Committing a Type II error is particularly costly for lower-power favor recipients who are disproportionately vulnerable to threats from higher-power favor givers (Keltner et al., 2003). In addition, higher-power favor givers may hold normative expectations that they should receive more grati-tude from lower-power favor recipients than vice versa. Thus, Type II errors should be more costly to and commonly experienced by lower-power favor recipients compared to higher-power favor recipients.

The consequences of a lower-power favor recipient com-mitting a Type I error, however, are comparatively trivial (e.g., the higher-power favor giver may feel awkwardly flat-tered). As such, a prudent interpersonal strategy for lower-power favor recipients to adopt when interacting with higher-power favor givers is to err on the side of expressing too much rather than too little gratitude even as the relational history between the two parties deepens. In other words, lower-power favor recipients may be wise to intentionally bias their interpersonal strategy toward committing rela-tively more Type I gratitude errors than Type II gratitude errors when interacting with higher power favor givers.

However, committing a Type II error is inherently less threatening for low-power favor recipients interacting with similarly low-power favor givers, who do not wield the threat of punishment. In this situation, the costs of expressing too much or too little gratitude are relatively low, and there-fore, the low-power favor recipient may determine that it is appropriate to express less gratitude as the relational history between the two low-power actors deepens and individuat-ing information about the favor giver confirms that the favor giver is indeed not a threat (Fiske & Neuberg, 1990).

In summary, we propose that low-power favor recipients will express a high level of gratitude when interacting with higher-power favor givers regardless of the length of their relational history, but will adjust their level of gratitude expression downward over the course of their relationships

with similarly low-power favor givers as incoming individu-ating information confirms the favor giver is not threatening. High-power favor recipients, on the other hand, will be less discerning of and concerned with the relational histories they have with specific favor givers as a function of those indi-viduals’ power level. As a result, the extent to which high-power (vs. low-power) favor recipients adjust their level of gratitude expression downward as their relational histories with others deepen will depend less on the power level of the favor giver because fewer, if any, favor givers will have the authority to punish higher-power favor recipients based on the favor recipients’ level of gratitude expression. Overall, this proposed pattern of behavior reflects a three-way interaction:

Hypothesis 5: A three-way interaction among favor recipient power, favor giver power, and relational history will emerge such that as the number of previous interac-tions with the same favor giver increases, lower-power (vs. higher-power) favor recipients will adjust their level of gratitude expression downward to a greater extent, but only when communicating with similarly low-power favor givers. When interacting with higher-power favor givers, lower-power favor recipients will express a con-sistently high level of gratitude, regardless of relational history length.

Overview of Studies

Four main studies explore the effects of relative power on expressions and feelings of gratitude. Study 1 examined a unique form of public gratitude expression by analyzing acknowledgements published in academic articles. Study 2 was a pre-registered experiment involving live conversations online in which we assessed expressions and feelings of grat-itude. Study 3 was another pre-registered experiment and assessed two potential mediators of the effects of relative power on gratitude expressions and feelings: interpersonal orientation and psychological entitlement. Study 4 assessed relational history as a potential moderator of the effect of relative power on gratitude expression using conversational data from Wikipedia editors. Finally, we report four addi-tional studies, including two that directly manipulate the mediators, in the supplementary online material (SOM) doc-ument which provide additional support for our hypotheses (see Figure 1).

Study 1: An Archival Analysis of Professional Power and Gratitude Expression

In Study 1, we tested Hypothesis 1 using a measure of objective power and the amount of gratitude expressed in the acknowledgements section of published academic articles.

Anicich et al. 5

Method

We acquired records of every article published in the Academy of Management Review (AMR), from the Business Source Complete and Web of Science databases over 40 years (1976–2015).1

Sample details. We excluded special contributions (e.g., Dia-logues, Replies, Editor’s Notes), articles that had at least one author for whom our power classification did not apply (e.g., authors identified as being from industry, Readers, Lecturers, etc.), and outliers (n = 6).2 Our final sample consisted of 1,272 articles.

Power measure. We utilized published biographical informa-tion to assess the authors’ level of organizational power (coded 1 = Graduate Student, 2 = Postdoctoral Fellow, 3 = Assistant Professor, 4 = Associate Professor, 5 = Full Pro-fessor; see Shane & Stuart, 2002). We created three measures of author power. First, we used first-author power only. Sec-ond, we used average author power, calculated by summing the individual author ranks and dividing by the number of authors at the article level. Third, we used a binary measure, tenured professor (coded 1 = first author is an Associate Professor or Full Professor, 0 = Otherwise). Past work has treated academic rank as a binary variable in this way (Aldridge et al., 2014; Libaers, 2014). Although we are unable to directly assess relative power in this setting, we believe that authors’ professional rank is a useful proxy for relative power (e.g., full professors predominantly interact with lower-power others while graduate students predominantly interact with higher-power others). This view is consistent with Anicich and Hirsh’s (2017a, p. 659, 2017b; Anicich, 2016) claim that power can be conceptualized “in terms of the stability of one’s vertical orientation” based on the “ratio of upward to downward interactions that one is likely to experi-ence, given the composition of one’s social network” (p. 662).

Gratitude expression. We assessed gratitude expression in three ways. First, two independent coders counted

the number of people and groups thanked in each article. Individuals thanked by name were counted as “people,” whereas collections of individuals (e.g., lab groups, seminar participants, and conference talk attendees, etc.) were counted as “groups.” The first author resolved coding dis-crepancies (<1% of articles). Second, we constructed a binary variable indicating whether or not the author(s) expressed any gratitude (i.e., thanked at least one person or group; coded 1 = expressed some gratitude, 0 = did not express any gratitude). Finally, we used the Linguistic Inquiry and Word Count (LIWC; Pennebaker et al., 2015) computerized text-analysis program to assess gratitude expression. LIWC leverages a dictionary of more than 2,300 words and word stems to identify the relative frequency with which particular word categories are used. We created a gratitude dictionary that included the following words: gratef*, grati*, thank*, or appreciat*. The “*” symbol means that all words that start with the same letters are included. For example, thank* includes “thanks,” “thanking,” “thank-ful,” and so on.

Control variables. We controlled for variables that could con-ceivably affect our predictor or outcome measures including the number of authors, the number of unique schools that the author group was affiliated with, article length (i.e., number of pages), the number of citations,3 the year of publication, author gender, and the word count of the acknowledgements section (when using the LIWC measure of gratitude as the dependent variable; see Table 1).

Results

All of the results reported below involved fixed-effects mod-els with journal issue as the grouping variable. Negative binomial regressions revealed that author power was nega-tively associated with the number of people and groups thanked using the first author power measure, b = −0.13, SE = 0.03, p < .001, 95% CI = [−0.18, −0.08], average author power measure, b = −0.24, SE = 0.03, p < .001, 95% CI = [−0.31, −0.17], and first author tenured measure (1 = yes, 0 = no), b = −0.24, SE = 0.06, p < .001, 95% CI = [−0.35, −0.13] (top portion of Table 2). Furthermore, logistic regres-sions revealed that author power was negatively associated with the likelihood of expressing any gratitude using the first author power measure, b = −0.40, SE = 0.08, p < .001, 95% CI = [−0.56, −0.25], the average author power measure, b = −0.61, SE = 0.10, p < .001, 95% CI = [−0.80, −0.42], and the first author tenured measure (1 = yes, 0 = no), b = −0.66, SE = 0.15, p < .001, 95% CI = [−0.95, −0.36] (middle portion of Table 2). Finally, linear regressions revealed that author power was negatively associated with gratitude expression using the LIWC gratitude dictionary using the first author power measure, b = −0.24, SE = 0.06, p < .001, 95% CI = [−0.36, −0.13], the average author power measure, b = −0.36, SE = 0.07, p < .001, 95% CI =

Figure 1. Model of tested relationships.

6

Tab

le 1

. D

escr

iptiv

e St

atis

tics

for

and

Cor

rela

tions

Am

ong

All

Inde

pend

ent,

Dep

ende

nt, a

nd C

ontr

ol V

aria

bles

(St

udy

1).

No

Var

iabl

eM

ean

SD1

23

45

67

89

1011

1213

1N

umbe

r of

Peo

ple

and

Gro

ups

Tha

nked

4.62

5.10

2A

ny G

ratit

ude

Expr

esse

d (1

= Y

es, 0

= N

o)0.

690.

460.

61**

3LI

WC

Mea

sure

of

Gra

titud

e Ex

pres

sion

2.14

2.07

0.26

**0.

69**

4Fi

rst

Aut

hor

Pow

er3.

631.

05–0

.09*

*–0

.13*

*–0

.11*

*

5A

vera

ge P

ower

3.68

0.86

–0.0

9**

–0.1

5**

–0.1

2**

0.73

**

6Fi

rst

Aut

hor

Ten

ured

(1

= y

es, 0

= n

o)0.

530.

50–0

.09*

*–0

.12*

*–0

.10*

*0.

82**

0.62

**

7N

umbe

r of

Aut

hors

1.83

0.81

0.06

*0.

050.

06*

0.02

0.05

0.06

*

8U

niqu

e Sc

hool

s1.

530.

720.

09**

0.05

0.07

*0.

10**

0.15

**0.

050.

72**

9

Art

icle

Len

gth

16.0

77.

120.

49**

0.37

**0.

23**

–0.0

20.

010.

000.

19**

0.16

**

10N

umbe

r of

Cita

tions

87.7

019

3.13

0.12

**0.

11**

0.04

0.00

–0.0

3–0

.03

0.01

0.00

0.24

**

11Y

ear

of P

ublic

atio

n19

92.9

511

.67

0.56

**0.

41**

0.29

**–0

.02

0.06

*–0

.02

0.33

**0.

33**

0.58

**0.

03

12A

vera

ge A

utho

r G

ende

r

(1 =

mal

e, 0

= fe

mal

e)0.

760.

37–0

.21*

*–0

.19*

*–0

.15*

*0.

12**

0.16

**0.

11**

–0.0

7*–0

.06*

–0.2

3**

–0.0

7**

–0.2

7**

13G

ende

r of

Fir

st A

utho

r

(1 =

mal

e, 0

= fe

mal

e)0.

760.

43–0

.18*

*–0

.17*

*–0

.13*

*0.

15**

0.12

**0.

13**

–0.0

7*–0

.06*

–0.2

0**

–0.0

6*–0

.25*

*0.

86**

14W

ord

Cou

nt o

f A

ckno

wle

dgm

ent

32.3

729

.19

0.78

**0.

68**

0.22

**–0

.06*

–0.0

9**

–0.0

50.

050.

06*

0.39

**0.

09**

0.45

**–0

.17*

*–0

.14*

*

Not

e. N

= 1

,272

. LIW

C =

Lin

guis

tic In

quir

y an

d W

ord

Cou

nt.

*p <

.05.

**p

< .0

1.

7

Tab

le 2

. R

egre

ssio

n R

esul

ts F

rom

Stu

dy 1

.

Dep

ende

nt v

aria

ble =

Num

ber

of p

eopl

e an

d gr

oups

tha

nked

Fixe

d-ef

fect

s ne

gativ

e bi

nom

ial r

egre

ssio

n (w

ith jo

urna

l iss

ue a

s th

e gr

oupi

ng v

aria

ble)

Firs

t A

utho

r Po

wer

–0.1

3***

(0.0

3)A

vera

ge A

utho

r Po

wer

–0.2

4***

(0.0

3)Fi

rst

Aut

hor

Ten

ured

(1 =

yes

, 0 =

no)

–0.2

4***

(0.0

6)N

umbe

r of

Aut

hors

–0.1

6**

(0.0

5)N

umbe

r of

Aut

hors

–0.1

7**

(0.0

5)N

umbe

r of

Aut

hors

–0.1

3*(0

.05)

Uni

que

Scho

ols

0.02

(0.0

6)U

niqu

e Sc

hool

s0.

06(0

.06)

Uni

que

Scho

ols

–0.0

1(0

.06)

Art

icle

Len

gth

0.05

***

(0.0

1)A

rtic

le L

engt

h0.

05**

*(0

.01)

Art

icle

Len

gth

0.05

***

(0.0

1)N

umbe

r of

Cita

tions

0.00

(0.0

0)N

umbe

r of

Cita

tions

0.00

(0.0

0)N

umbe

r of

Cita

tions

0.00

(0.0

0)A

utho

r G

ende

r–0

.08

(0.0

6)A

utho

r G

ende

r–0

.11

(0.0

8)A

utho

r G

ende

r–0

.10

(0.0

6)C

onst

ant

0.08

(0.1

7)C

onst

ant

0.51

**(0

.19)

Con

stan

t–0

.26

(0.1

6)

Dep

ende

nt v

aria

ble =

Whe

ther

or

not

any

peop

le o

r gr

oups

wer

e th

anke

d (1

= y

es, 0

= n

o)

Var

iabl

eC

oeff.

SEV

aria

ble

Coe

ff.SE

Var

iabl

eC

oeff.

SE

Fixe

d-ef

fect

s lo

gist

ic r

egre

ssio

n (w

ith jo

urna

l iss

ue a

s th

e gr

oupi

ng v

aria

ble)

Firs

t A

utho

r Po

wer

–0.4

0***

(0.0

8)A

vera

ge A

utho

r Po

wer

–0.6

1***

(0.1

0)Fi

rst

Aut

hor

Ten

ured

(1 =

yes

, 0 =

no)

–0.6

6***

(0.1

5)N

umbe

r of

Aut

hors

–0.3

2*(0

.14)

Num

ber

of A

utho

rs–0

.35*

(0.1

4)N

umbe

r of

Aut

hors

–0.2

5(0

.14)

Uni

que

Scho

ols

–0.1

9(0

.15)

Uni

que

Scho

ols

–0.1

3(0

.15)

Uni

que

Scho

ols

–0.2

5(0

.15)

Art

icle

Len

gth

0.10

***

(0.0

2)A

rtic

le L

engt

h0.

10**

*(0

.02)

Art

icle

Len

gth

0.10

***

(0.0

2)N

umbe

r of

Cita

tions

0.00

(0.0

0)N

umbe

r of

Cita

tions

0.00

(0.0

0)N

umbe

r of

Cita

tions

0.00

(0.0

0)A

utho

r G

ende

r–0

.29

(0.2

1)A

utho

r G

ende

r–0

.34

(0.2

6)A

utho

r G

ende

r–0

.33

(0.2

1)

Dep

ende

nt v

aria

ble =

LIW

C m

easu

re o

f gra

titud

e ex

pres

sion

Fixe

d-ef

fect

s lin

ear

regr

essi

on (

with

jour

nal i

ssue

as

the

grou

ping

var

iabl

e)

Firs

t A

utho

r Po

wer

–0.2

4***

(0.0

6)A

vera

ge A

utho

r Po

wer

–0.3

6***

(0.0

7)Fi

rst

Aut

hor

Ten

ured

(1 =

yes

, 0 =

no)

–0.4

4***

(0.1

2)N

umbe

r of

Aut

hors

–0.0

9(0

.10)

Num

ber

of A

utho

rs–0

.11

(0.1

1)N

umbe

r of

Aut

hors

–0.0

4(0

.11)

Uni

que

Scho

ols

0.00

(0.1

2)U

niqu

e Sc

hool

s0.

05(0

.12)

Uni

que

Scho

ols

–0.0

5(0

.12)

Art

icle

Len

gth

0.04

*(0

.02)

Art

icle

Len

gth

0.03

*(0

.02)

Art

icle

Len

gth

0.04

*(0

.02)

Num

ber

of C

itatio

ns–0

.00

(0.0

0)N

umbe

r of

Cita

tions

–0.0

0(0

.00)

Num

ber

of C

itatio

ns–0

.00

(0.0

0)A

utho

r G

ende

r–0

.21

(0.1

5)A

utho

r G

ende

r–0

.23

(0.1

7)A

utho

r G

ende

r–0

.23

(0.1

5)W

ord

Cou

nt0.

01**

(0.0

0)W

ord

Cou

nt0.

01**

(0.0

0)W

ord

Cou

nt0.

01**

(0.0

0)C

onst

ant

2.53

***

(0.3

6)C

onst

ant

3.02

***

(0.4

0)C

onst

ant

1.90

***

(0.3

2)

Not

e. N

= 1

,025

in a

ll lo

git

mod

els,

N =

1,2

72 in

all

othe

r m

odel

s.Fi

rst

auth

or g

ende

r w

as u

sed

to c

alcu

late

the

aut

hor

gend

er c

ontr

ol v

aria

ble

whe

n us

ing

the

first

aut

hor

pow

er m

easu

re a

nd a

vera

ge a

utho

r ge

nder

was

use

d to

cal

cula

te t

he a

utho

r ge

nder

con

trol

va

riab

le w

hen

usin

g th

e av

erag

e au

thor

pow

er m

easu

re (

code

d 1 =

mal

e, 0

= fe

mal

e). T

he a

naly

ses

reve

al t

hat

acro

ss a

ll m

easu

res

of p

ower

, gra

titud

e ex

pres

sion

was

neg

ativ

ely

corr

elat

ed w

ith a

utho

r po

wer

. SE =

sta

ndar

d er

ror;

LIW

C =

Lin

guis

tic In

quir

y an

d W

ord

Cou

nt.

*p <

.05.

**p

< .0

1. *

**p <

.001

.

8 Personality and Social Psychology Bulletin 00(0)

[−0.50, −0.22], and the first author tenured measure (1 = yes, 0 = no), b = −0.44, SE = 0.12, p < .001, 95% CI = [−0.67, −0.20] (bottom portion of Table 2).

Discussion

Lower-power authors compared to higher-power authors expressed more gratitude in the acknowledgements section of published academic articles, supporting Hypothesis 1. A limitation is that lower-power researchers may have received more help from others than higher-power research-ers, which could explain this effect. However, this rationale may not hold in academia. First, higher-power scholars are more likely than lower-power scholars to give external talks where they may receive helpful feedback. Second, higher-power scholars have larger networks from which to solicit feedback than lower-power scholars. Third, if a lower-power scholar receives substantial help on a manuscript, it may be expected that the helper be given authorship and hence not be thanked in the acknowledgements. We address this limita-tion by manipulating power and holding the magnitude and objective helpfulness of the favor constant in Studies 2 and 3.

Study 2: Experimental Evidence of Relative Power’s Effect on Expressions and Feelings of Gratitude During a Live Interaction

Study 2 was a pre-registered experiment that tested Hypotheses 1 and 2 (pre-registration: https://aspredicted.org/blind.php?x=bn4qh7). Specifically, we manipulated relative power using actual and asymmetric control over an objective resource, held constant the objective helpfulness of the favor, and measured gratitude expression using a behavioral mea-sure and felt gratitude using a paradigm involving a live online chat. We also measured the favor recipient’s percep-tion of how costly the favor was for the favor giver and how valuable the favor recipient perceived the favor to be and predicted that no differences would emerge between the two conditions on these measures because we intentionally held the objective cost/value of the favor constant across condi-tions. We note, however, that we only pre-registered our null hypothesis related to favor cost. Additional robustness checks and exploratory analyses are available in the SOM document.

Method

Participants. Two hundred and eighty-three Mechanical Turk workers participated in exchange for US$0.40. We overs-ampled to acquire at least 240 responses after applying our exclusion criteria (for similar approaches, see Anicich et al., 2020; Gan et al., 2018), which an a priori power analysis revealed was the necessary sample size to detect an effect

with 80% power (two-tailed, alpha = .05) based on the results of Study S1 reported in the SOM document. As speci-fied in our pre-registration document, we excluded partici-pants who failed an attention check (N = 14) and who expressed suspicion about their chat partner (N = 8), result-ing in a final sample of 261 (Mage = 37.74 years, SDage = 11.33, 59% female).

Relative power manipulation. We told participants to imagine working in an organization with employees distributed across the country. We further told participants that they may have the opportunity to interact with the other participant who was also ostensibly participating in the study. In reality, one of the current authors served as the confederate during the chat and followed a standardized script (described below).

We informed participants that they would have a chance to be entered into a virtual lottery for a US$25 Amazon gift card that would be paid out after all data were collected and that their chance of winning would be based on their number of lottery tickets.

We manipulated relative power by randomly assigning participants to either a lower-power condition (i.e., subordi-nate role) or a higher-power condition (i.e., boss role; see Waytz et al., 2015). Specifically, participants read,

The Boss gets to make the decision of how to divide up the 10 lottery tickets any way he/she likes. The Boss can keep all the lottery tickets for himself/herself, can split the lottery tickets equally, or give more or less to the other person. The Boss will keep the lottery tickets he/she allocates to himself/herself. The Subordinate does not get to make any decision. He/she simply receives the number of lottery tickets allocated by the Boss.

On the subsequent screen, participants in the lower-power [higher-power] condition read,

You have been assigned to the role of SUBORDINATE [BOSS]. As the subordinate [boss], you have far less [more] power than your boss [subordinate] who is also participating in this study. At the end of the study, your boss [you] will decide how many of the 10 lottery tickets to allocate to themselves and you [yourself and your subordinate].

Next, we told participants in both conditions that they would now complete several tasks as part of the organiza-tional simulation. Then, lower-power [higher-power] par-ticipants read the following task description adapted from Schmeichel (2007):

First, imagine that the organization you work for highly values attention to detail and creativity. To demonstrate your attention to detail and creativity, please write two pages about your day yesterday or as much as you are able to write in 5 minutes. However, there is an important rule—You may not use words containing the letters “A” or “N.” During this task, your boss [subordinate] will be given the option to chat with you. Your boss [subordinate] may or may not choose to initiate a chat.

Anicich et al. 9

Participants then saw two blank text entry boxes next to a timer with a reminder to write for at least 5 minutes. After 14 seconds, however, the screen automatically advanced and revealed a mail icon with the note: “Your boss [subordi-nate] has requested the opportunity to chat with you.” Then, a ChatPlat window appeared (e.g., see Brown et al., 2020; Huang et al., 2017) and participants read that they were now connected with their boss [subordinate]. Participants had 2 min to chat.

As we described in our pre-registration document, one of the authors played the role of a confederate (either the boss or subordinate role) and sent only two messages to participants:

Statement #1: “Hi there”

Statement #2 (after a brief pause): “I read about the writing task you have to complete. As your subordinate [boss], I am willing to complete the task for you . . . so just click to the next screen when you’re ready. ive already accepted the task on my end. ok?”

Gratitude expression. We measured gratitude expression based on whether the participant (i.e., the favor recipient) expressed any gratitude to the ostensibly higher-power or lower-power confederate (i.e., favor giver) by using any LIWC gratitude dictionary word from Study 1. We also assessed gratitude expression as a continuous variable using the LIWC gratitude dictionary.

Felt gratitude. After the chat, participants completed a 3-item measure of felt gratitude adapted from MacKenzie and Bau-meister (2019; e.g., “I feel extremely appreciative for the favor”; α = .98).

Perceived cost to the favor giver. Four items (α = .88) adapted from MacKenzie and Baumeister (2019) assessed how costly participants perceived the favor to be for their boss [subordi-nate] (e.g., “It cost this person a great deal of effort to do this favor for me”; 1 = “agree a little” to 9 = “very strongly agree”).

Perceived favor value. Five items (α = .97) adapted from MacKenzie and Baumeister (2019) assessed how valuable the favor recipient perceived the favor to be (e.g., “I consider this to be a very big favor”; 1 = “agree a little” to 9 = “very strongly agree”).

Manipulation check. To determine whether our manipulation was effective, participants responded to three 7-point response scales (α = .95) with the following response anchors: “I felt less [more] powerful than the other person,” “I had less [more] control than the other person,” and “I felt weaker [stronger] than the other person.”

Afterwards, participants were fully debriefed and assigned an equal chance of winning the US$25 lottery, which we paid out at the conclusion of the study.

Results

Our manipulation was effective. Lower-power participants (M = 2.46, SD = 1.37) reported having less power relative to their interaction partner than higher-power participants (M = 4.83, SD = 1.52), t(259) = 13.20, p < .001, d = 1.63.

Supporting Hypothesis 1, 40% of lower-power and 26% of higher-power favor recipients expressed gratitude, χ2(1, N = 261) = 6.01, p = .014. In addition, lower-power favor recipients (M = 7.55, SD = 11.24) expressed signifi-cantly more gratitude than higher-power favor recipients using the LIWC dictionary measure (M = 3.98, SD = 8.36), t(259) = 2.91, p = .004, d = 0.36. Supporting Hypothesis 2, lower-power favor recipients (M = 7.04, SD = 2.13) reported feeling more gratitude than higher-power favor recipients (M = 6.06, SD = 2.66), t(259) = 3.29, p = .001, d = 0.41.

Finally, our predictions that participants in the two condi-tions would report no differences in the perceived cost of the favor to the favor giver and perceived value of the favor to the favor recipient were not supported. Participants in the lower-power favor recipient condition (M = 5.58, SD = 2.23) reported perceiving the favor as more costly to the favor giver than participants in the higher-power favor recip-ient condition (M = 4.72, SD = 2.19), t(259) = 3.17, p = .002, d = 0.39. We also found that participants in the lower-power favor recipient condition (M = 6.70, SD = 2.19) reported perceiving the favor as more valuable than partici-pants in the higher-power favor recipient condition (M = 5.90, SD = 2.39), t(259) = 2.83, p = .005, d = 0.35. One explanation that may account for these unexpected findings is that individuals may, on average, believe that higher-power favor givers’ time is more valuable than lower-power favor givers’ time. To the extent this is a widely held belief, when higher-power compared to lower-power individuals perform a favor, the favor recipient may perceive that favor to be both more costly to the favor giver and more valuable to the self.

Since relative power was significantly associated with the perceived cost of the favor for the favor giver and the per-ceived value of the favor to the self, we re-ran our analyses while controlling for these variables. Relative power remained significantly and negatively associated with any gratitude expression, logistic regression: b = −0.55, SE = 0.27, p = .043, and was marginally significantly and nega-tively associated with felt gratitude, linear regression: b = −0.28, SE = 0.16, p = .083.

Discussion

Study 2 provided support for Hypotheses 1 and 2 that power reduces expressions and feelings of gratitude in an

10 Personality and Social Psychology Bulletin 00(0)

objectively helpful favor situation. These effects remained significant or marginally significant when controlling for the perceived favor cost and value.

Study 3: Experimental Evidence of Mediation Through Interpersonal Orientation and Psychological Entitlement

Study 3 was another pre-registered experiment in which participants indicated how much gratitude they would express and feel after receiving a favor in an organizational scenario (pre-registration: https://aspredicted.org/blind.php?x=p6y527). We randomly assigned participants to one of three relative power conditions (i.e., lower, same, or higher) and measured our proposed mediators―interper-sonal orientation and psychological entitlement. Thus, Study 3 tested Hypotheses 1–4b.

Method

Participants. Three-hundred and fifty-six Mechanical Turk workers participated in exchange for US$0.40. We overs-ampled to acquire at least 339 responses after applying our exclusion criteria, which an a priori power analysis revealed was the necessary sample size to detect a small to medium effect (i.e., Cohen’s f = 0.17, α = .05) based on the results of Study S2 reported in the SOM document. We excluded six participants who failed an attention check, resulting in a final sample of 350 (Mage = 38.44 years, SDage = 12.81 years; 57% female).

Power manipulation. Participants read one of three organiza-tional scenarios (adapted from Blader & Chen, 2012; Mooij-man et al., 2020). Participants in the relatively higher-power [lower-power] condition read,

Imagine that you are an employee in a medium-sized firm. Further, imagine having hierarchical relationships with others in your organization. This means that you hold more [less] power than other employees and you control many more [fewer] resources compared to other employees in the organization. Indeed, you are one of the most [least] powerful individuals in the organization. Other employees report directly to you [You directly report to other employees].

Participants in the same-power condition read,

Imagine that you are an employee in a medium-sized firm. Further, imagine having non-hierarchical relationships with others in your organization. This means that you hold the same amount of power as other employees and you control an equal amount of resources compared to other employees in the organization. Indeed, you are neither more nor less powerful

than most individuals in the organization. Neither you nor the other employees directly report to each other.

We did not specify the participant’s absolute level of power in the same-power condition to ensure our manipula-tion did not simultaneously manipulate relative and absolute power, which would have introduced a confound.

Interpersonal orientation (mediator). We assessed interpersonal orientation using six items (α = .79, 1 = “strongly disagree” to 7 = “strongly agree”), adapted from three related mea-sures. Our measure included two items from the IO Scale (Swap & Rubin, 1983): “I would be greatly influenced by the moods of others in the organization” and “I would be very sensitive to criticism from others in the organization,” two items from Hashimoto and Yamagishi’s (2013) measure of interdependence: “I would find myself being concerned about what others in the organization think of me” and “I would think it is important not to disturb good relations with others in the organization,” and two items from the Interdependent Self-Construal Scale (Singelis, 1994): “Even if I strongly dis-agreed with others in the organization, I would avoid an argu-ment,” and “I would feel that my fate is intertwined with the fate of others in the organization.”

Psychological entitlement (mediator). Four items (α = .91) measured psychological entitlement (adapted from Camp-bell et al., 2004): “I would honestly feel I’m just more deserving than others in the organization,” “I would feel entitled to better outcomes than others in the organization,” “Things should go my way more than they should go the way of others in the organization,” and “I would deserve more things in my life than others in the organization” (1 = “strongly disagree” to 7 = “strongly agree”). We counterbal-anced the order of the interpersonal orientation and psycho-logical entitlement measures.

Gratitude expression. Next, participants read, “Now imagine that you received the following email from your subordinate [coworker] [boss], Sam.” The email read,

Hi [participant’s first name inserted here],

Following up on our conversation from the other day, I wanted to share this article I came across. As you can see, it’s highly relevant to what we were discussing.

Best,

Sam

Participants then read, “It turns out that you had recently heard about this article and were looking forward to reading it, but had trouble finding it online,” which we included to indicate the favor was actually helpful. Finally, participants

Anicich et al. 11

selected one of the following ordinal responses to send to Sam: (1) “[no response],” (2) “Interesting, I’ll have a look,” (3) “Interesting, I’ll have a look. Thanks,” and (4) “Interesting, I’ll have a look. Thanks so much for sending this!”

Felt gratitude. We measured felt gratitude with the same three items from the previous study, but worded as hypothetical statements (e.g., “I would feel extremely appreciative for the favor”; α = .93).

Manipulation check. Participants responded to the same 3-item measure we used in the previous study (α = .96).

Results

Our manipulation was successful. Higher-power participants (M = 5.16, SD = 1.15) reported having more power than same-power participants (M = 4.04, SD = 0.62), t(231) = 9.28, p < .001, d = 1.21, who reported having more power than lower-power participants (M = 2.77, SD = 1.40), t(232) = 8.92, p < .001, d = 1.17.

Expressed gratitude. To test Hypothesis 1, we conducted Mann–Whitney U-tests, which is appropriate when com-paring differences between two independent groups when the dependent variable is ordinal, but not normally distrib-uted. Higher-power favor recipients (M = 3.30, Mdn = 3.0, SD = 0.69) expressed significantly less gratitude than lower-power favor recipients (M = 3.64, Mdn = 4.0, SD = 0.52), U = 4,955.5, z = 4.04, p < .001, and same-power favor recipients (M = 3.45, Mdn = 4.0, SD = 0.74), U = 5,785.0, z = 2.18, p = .029). Same-power favor recipients expressed marginally less gratitude than lower-power favor recipients, U = 6,049.5, z = 1.80, p = .073).

Felt grat itude. To test Hypothesis 2, we conducted independent-samples t-tests. Higher-power favor recipients (M = 5.84, SD = 1.94) did not feel significantly less grateful than either lower-power favor recipients (M = 6.09, SD =

Table 3. Regression Results Pertaining to Mediation for Study 3.

Variable

DV = Interpersonal orientation

DV = Psychological entitlement

DV = Expressed gratitude

DV = Expressed gratitude

DV = Felt gratitude

DV = Felt gratitude

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Coeff. SE SE Coeff SE Coeff. SE Coeff. SE Coeff. SE

Power (–1 = LP, 0 = SP, 1 = HP)

–0.28*** (0.06) 0.29** (0.09) –0.17*** (0.04) –0.13** (0.04) –0.12 (0.12) 0.02 (0.12)

Interpersonal Orientation

0.07* (0.04) 0.39*** (0.10)

Psychological Entitlement

–0.05* (0.03) –0.14+ (0.07)

Note. N = 350 in all models. SE = standard error.*p < .05. **p < .01. ***p < .001. +p = .06.

1.92), t(231) = 0.99, p = .33, d = 0.13, or same-power favor recipients (M = 6.17, SD = 1.54), t(231) = 1.46, p = .15, d = 0.19. Same-power and lower-power favor recipients did not differ in their levels of felt gratitude, t(232) = 0.38, p = .71, d = 0.05. Thus, we did not find support for Hypothesis 2.

Interpersonal orientation. As predicted, lower-power favor recipients (M = 5.00, SD = 1.10) reported greater interper-sonal orientation than higher-power favor recipients (M = 4.45, SD = 0.95), t(231) = 4.10, p < .001, d = 0.54. Lower-power favor recipients also reported greater interpersonal orientation than same same-power favor recipients, t(232) = 3.54, p < .001, d = 0.46. Unexpectedly, however, higher-power favor recipients and same-power favor recipients (M = 4.53, SD = 0.91) did not differ in their reported level of IO, t(231) = 0.70, p = .49, d = 0.09.

Psychological entitlement. As predicted, higher-power favor recipients (M = 3.42, SD = 1.39) reported more psychologi-cal entitlement than lower-power favor recipients (M = 2.85, SD = 1.24), t(231) = 3.33, p = .001, d = 0.44, and same-power favor recipients (M = 2.81, SD = 1.32), t(231) = 3.47, p = .001, d = 0.45. Same-power and lower-power favor recipients did not differ in psychological entitlement, t(232) = 0.24, p = .81, d = 0.03.

Mediation analysis. As outlined in our pre-registration docu-ment, we coded relative power as a continuous variable (coded −1 = lower-power condition, 0 = same-power con-dition, 1 = higher-power condition) and tested our media-tion hypotheses. A bootstrapping procedure with 5,000 resamples (Hayes, 2013) indicated that interpersonal orien-tation (95% CI = [−0.047, −0.001]) and psychological enti-tlement (95% CI = [−0.038, −0.001]) simultaneously and partially mediated the effect of relative power on gratitude expression, supporting Hypotheses 3a and 4a.

When felt gratitude was the outcome, we conducted indi-rect-only mediation (see Zhao et al., 2010) because the direct

12 Personality and Social Psychology Bulletin 00(0)

effect of relative power on felt gratitude (i.e., the c path) was not significant. A bootstrapping procedure with 5,000 resa-mples indicated that interpersonal orientation significantly and fully mediated the effect of relative power on felt grati-tude (95% CI = [−0.206, −0.033]), but psychological entitle-ment did not function as a significant mediator (95% CI = [−0.103, 0.005]), providing support for Hypothesis 4b, but not Hypothesis 4a (see Table 3).

To provide greater confidence in our proposed mecha-nisms, we report two additional experiments involving an objectively helpful favor in the SOM document. Following the experimental-causal-chain approach (Spencer et al., 2005), we find that participants in a high (vs. low) interper-sonal orientation condition expressed and felt more gratitude (Supplemental Study S3a). In addition, participants in a high (vs. low) psychological entitlement condition expressed and felt less gratitude (Supplemental Study S3b). These results provide additional evidence that relative power influences feelings and expressions of gratitude through differences in interpersonal orientation and psychological entitlement.

Discussion

Study 3 provided additional evidence that lower-power (compared to higher-power) favor recipients express more gratitude. However, unlike the results of Study 2, relative power was not significantly associated with feelings of grati-tude. Furthermore, we demonstrated that the effect of rela-tive power on gratitude expression was simultaneously mediated by lower-power individuals experiencing increased interpersonal orientation compared to higher-power individ-uals and higher-power individuals experiencing increased psychological entitlement compared to lower-power indi-viduals. With respect to feelings of gratitude, only IO, but not psychological entitlement, significantly mediated the effect.

Study 4: Conversational Exchanges Among Wikipedia Editors

In Study 4, we leveraged a rich, archival dataset of conversa-tional exchanges among Wikipedia editors (N = 136,215). Because the dataset includes information about the power of the favor recipient and favor giver, we were able to assess the effect of relative power on gratitude expression (Hypothesis 1). Furthermore, given the panel structure of our dataset, we were able to test the three-way interaction among favor recipient power, favor giver power, and relational history (Hypothesis 5).

Method

We acquired detailed conversational records from Wikipedia editors.4 The final sample included 136,215 observations (12,681 unique speakers) from 2003 to 2011. Each observa-tion is a comment submitted by one editor to another editor.

These conversations occurred on talk pages, which are forums where Wikipedia editors discuss improvements to articles. Additional study details are provided in the SOM document.

Favor recipient power. Favor recipient power was a dummy variable that distinguishes between the two types of Wikipedia editors—administrators (or “admins”) and non-admin editors. Admin editors have more formal power than non-admin editors because admins have unique access to “restricted technical features such as protecting or delet-ing pages or blocking other editors” (Danescu-Niculescu-Mizil et al., 2012, p. 701), responsibilities which are consistent with our definition of power (Magee & Galin-sky, 2008).

Relational history (moderator). We operationalized relational history as the number of comments the favor recipient had previously sent to a particular favor giver.

Gratitude expression. We assessed gratitude expression using the LIWC gratitude dictionary from Studies 1 and 2 (i.e., gratef*, grati*, thank*, or appreciat*). In the SOM docu-ment, we also report consistent results when using whether or not speakers expressed gratitude as a binary measure (coded 1 = expressed gratitude, 0 = did not express any gratitude).

Control variables. We controlled for speaker gender (coded 0 = female, 1 = male, and 2 = unknown), recipient gender (0 = female, coded 1 = male, and 2 = unknown), speaker edit count (i.e., logged number of speaker’s previous Wiki-pedia edits), recipient edit count (i.e., logged number of recipient’s previous Wikipedia edits), comment word count, and preceding comment word count (i.e., of the comment to which the current speaker is responding). We included fixed effects to control for the comment year (see Table 4).

Results

We ran mixed effects regressions with favor recipient ID as the grouping variable to account for the fact that comments were nested within speakers (N = 12,681 favor recipients and 136,215 comments).

The interaction between favor recipient power and favor giver power on gratitude expression was significant, b = −0.55, SE = 0.12, p < .001, 95% CI = [−0.79, −0.31] (Model 1, Table 5). Specifically, high-power favor recipients expressed less gratitude to low-power favor givers (M = 1.52, SE = 0.04) than to high-power favor givers (M = 2.22, SE = 0.08), t(42, 371) = −8.09, p < .001, d = 0.08. In addi-tion, low-power favor recipients expressed more gratitude to high-power favor givers (M = 3.26, SE = 0.07) than to low-power favor givers (M = 2.05, SE = 0.04), t(93, 870) = −16.84, p < .001, d = 0.11 (see Figure 2).

13

Tab

le 4

. D

escr

iptiv

e St

atis

tics

and

Cor

rela

tions

Am

ong

All

Var

iabl

es U

sed

in S

tudy

4.

No.

Var

iabl

eM

ean

SD1

23

45

67

89

1011

1G

ratit

ude

Expr

essi

on

(LIW

C m

easu

re #

1)2.

199.

48

2G

ratit

ude

Expr

essi

on

(LIW

C m

easu

re #

2)2.

389.

500.

990*

*

3Fa

vor

Rec

ipie

nt P

ower

(i.

e., A

dmin

= 1

, N

on-A

dmin

Edi

tor =

0)

0.31

0.46

–0.0

32**

–0.0

32**

4Fa

vor

Giv

er P

ower

(i.e

., A

dmin

= 1

, N

on-A

dmin

Edi

tor =

0)

0.29

0.46

0.04

9**

0.04

9**

0.02

4**

5Fa

vor

Rec

ipie

nt E

dit

C

ount

(lo

gged

)9.

552.

06–0

.014

**–0

.015

**0.

260*

*–0

.031

**

6Fa

vor

Giv

er E

dit

Cou

nt (

logg

ed)

9.46

2.19

0.02

8**

0.02

6**

–0.0

15**

0.25

7**

0.01

3**

7

Favo

r R

ecip

ient

Gen

der

(1 =

mal

e,

2 =

fem

ale,

3 =

unk

now

n)2.

390.

910.

002

0.00

3–0

.036

**0.

022*

*–0

.169

**–0

.029

**

8Fa

vor

Giv

er G

ende

r (1

= m

ale,

2 =

fem

ale,

3 =

unk

now

n)2.

360.

92–0

.015

**–0

.013

**0.

010*

*–0

.064

**–0

.037

**–0

.193

**0.

052*

*

9W

ord

Cou

nt o

f Cur

rent

Com

men

t47

.38

51.0

4–0

.168

**–0

.171

**–0

.019

**–0

.028

**–0

.003

–0.0

13**

0.00

10.

012*

*

10W

ord

Cou

nt o

f Pre

cedi

ng C

omm

ent

29.6

614

.67

–0.0

81**

–0.0

81**

–0.0

25**

–0.0

07*

–0.0

11**

0.01

7**

0.00

40.

001

0.16

9**

11

Yea

r20

08.9

11.

830.

014*

*0.

012*

*–0

.059

**–0

.139

**0.

129*

*0.

112*

*–0

.203

**–0

.205

**0.

024*

*0.

011*

*

12R

elat

iona

l His

tory

(i.e

., no

. of

prev

ious

inte

ract

ions

)2.

9511

.72

–0.0

23**

–0.0

25**

0.05

0**

0.00

20.

074*

*0.

081*

*0.

027*

*0.

036*

*–0

.017

**–0

.017

**–0

.022

**

Not

e. N

= 1

36,2

15. L

IWC

= L

ingu

istic

Inqu

iry

and

Wor

d C

ount

.*p

< .0

5. *

*p <

.01.

14

Tab

le 5

. R

egre

ssio

n R

esul

ts fo

r St

udy

4.

Var

iabl

e

DV

= A

mou

nt o

f gra

titud

e ex

pres

sed

(LIW

C s

core

s)D

V =

Any

gra

titud

e ex

pres

sed

(L

IWC

sco

res,

1 =

Yes

, 0 =

No)

Mod

el 1

Mod

el 2

Mod

el 3

Mod

el 4

Coe

ff.SE

Coe

ffSE

Coe

ffSE

Coe

ffSE

Favo

r R

ecip

ient

Pow

er

(i.e.

, Adm

in =

1, N

on-A

dmin

Edi

tor =

0)

–0.5

94**

*(0

.125

)–0

.612

***

(0.1

26)

–0.3

24**

*(0

.031

)–0

.322

***

(0.0

32)

Favo

r G

iver

Pow

er

(i.e.

, Adm

in =

1, N

on-A

dmin

Edi

tor =

0)

1.11

6***

(0.0

73)

1.07

2***

(0.0

75)

0.38

7***

(0.0

20)

0.35

9***

(0.0

21)

Favo

r R

ecip

ient

Edi

t C

ount

(i.e

., lo

gged

)–0

.064

*(0

.030

)–0

.063

*(0

.030

)–0

.047

***

(0.0

06)

–0.0

45**

*(0

.006

)Fa

vor

Giv

er E

dit

Cou

nt (

i.e.,

logg

ed)

0.04

8***

(0.0

12)

0.05

4***

(0.0

12)

0.01

2**

(0.0

04)

0.01

6***

(0.0

04)

Favo

r R

ecip

ient

Gen

der

(ref

. fem

ale)

M

ale

–0.3

59(0

.546

)–0

.363

(0.5

46)

–0.2

60*

(0.1

04)

–0.2

70**

(0.1

04)

U

nkno

wn

–0.4

17(0

.537

)–0

.412

(0.5

37)

–0.2

99**

(0.1

02)

–0.3

05**

(0.1

02)

Favo

r G

iver

Gen

der

(ref

. fem

ale)

M

ale

0.20

4(0

.166

)0.

192

(0.1

66)

0.02

9(0

.048

)0.

019

(0.0

48)

U

nkno

wn

0.15

5(0

.163

)0.

157

(0.1

63)

0.03

5(0

.047

)0.

030

(0.0

47)

Wor

d C

ount

Cur

rent

Com

men

t–0

.029

***

(0.0

01)

–0.0

29**

*(0

.001

)–0

.003

***

(0.0

00)

–0.0

03**

*(0

.000

)W

ord

Cou

nt o

f Pre

cedi

ng C

omm

ent

–0.0

37**

*(0

.002

)–0

.037

***

(0.0

02)

–0.0

04**

*(0

.001

)–0

.004

***

(0.0

01)

Favo

r R

ecip

ient

Pow

er ×

Fav

or G

iver

Pow

er–0

.551

***

(0.1

20)

–0.4

89**

*(0

.126

)–0

.087

*(0

.035

)–0

.043

(0.0

37)

Fam

iliar

ity w

ith F

avor

Giv

er (

i.e.,

# o

f pre

viou

s in

tera

ctio

ns)

–0.0

31**

*(0

.006

)–0

.028

***

(0.0

02)

Favo

r R

ecip

ient

Pow

er ×

Len

gth

of C

onve

rsat

iona

l His

tory

0.02

1***

(0.0

06)

0.01

3***

(0.0

03)

Favo

r G

iver

Pow

er ×

Len

gth

of C

onve

rsat

iona

l His

tory

0.02

2**

(0.0

08)

0.01

9***

(0.0

03)

Favo

r R

ecip

ient

Pow

er ×

Fav

or G

iver

Pow

er ×

Len

gth

of

Con

vers

atio

nal H

isto

ry–0

.035

*(0

.015

)–0

.030

***

(0.0

06)

Yea

r Fi

xed

Effe

cts

Yes

Yes

Yes

Yes

C

onst

ant

4.77

6***

(1.2

40)

4.73

3***

(1.2

40)

–1.0

79**

(0.3

35)

–1.1

09**

(0.3

40)

Num

ber

of C

omm

ents

136,

215

136,

215

136,

215

136,

215

N

umbe

r of

Uni

que

Spea

kers

12,6

8112

,681

12,6

8112

,681

Not

e. M

odel

s 1

and

2 ar

e lin

ear

mix

ed e

ffect

s re

gres

sion

s (u

sing

the

xtr

eg c

omm

and

in S

tata

) an

d M

odel

s 3

and

4 (u

sing

the

glm

er c

omm

and

in R

). A

ll m

odel

s in

clud

ed fa

vor

reci

pien

t ID

as

the

grou

ping

va

riab

le. L

IWC

= L

ingu

istic

Inqu

iry

and

Wor

d C

ount

; SE =

sta

ndar

d er

ror.

*p <

.05.

**p

< .0

1. *

**p <

.001

.

Anicich et al. 15

Next, we tested whether relational history moderated the effect of relative power on gratitude expression. As pre-dicted, a three-way interaction among favor recipient power, favor giver power, and relational history on gratitude expres-sion emerged, b = −0.04, SE = 0.02, p = .02, 95% CI = [−0.066, −0.005] (Model 2, Table 5). Specifically, when the favor recipient had low power, the two-way interaction between favor giver power and relational history was signifi-cant, b = 0.022, SE = 0.008, p = .006, 95% CI = [0.006, 0.038] (left panel of Figure 3). However, when the favor recipient had high power, the two-way interaction between

favor giver power and relational history was not significant, b = −0.0003, SE = 0.0117, p = .98, 95% CI = [−0.023, 0.023] (right panel of Figure 3).

This pattern of results demonstrates that as relational history increased, low-power (vs. high-power) favor recipi-ents adjusted their level of gratitude expression downward, but only when communicating with similarly low-power favor givers. When interacting with higher-power favor givers, lower-power favor recipients expressed a consis-tently high level of gratitude, regardless of relational his-tory length.

Figure 2. Gratitude expression in Study 4. Error bars represent ±1 Standard Error.

Figure 3. Interaction between favor giver power and length of relational history by favor recipient power level in Study 4.

16 Personality and Social Psychology Bulletin 00(0)

Discussion

Study 4 used a large archival dataset of actual conversational exchanges and demonstrated that relational history moder-ated the effect of relative power on gratitude expression. As in Study 1, one limitation is that lower-power non-admin edi-tors may have received more help from higher-power admins, which could explain this effect. Although we are unable to rule out this alternative explanation, it is worth noting that admins and non-admin editors closely collaborate to edit Wikipedia content. Therefore, giving and receiving help is crucial to the successful maintenance of Wikipedia as a com-munal resource (e.g., see quotes in SOM Table S4 depicting the collaboration process). In addition, the experimental results we report in Studies 2 and 3 address this concern.

General Discussion

Power and gratitude are universal features of social life. We reported four main studies and four supplementary studies that established a negative relationship between relative power and expressing and feeling gratitude. Furthermore, we identified two mechanisms linking relative power to grat-itude—interpersonal orientation and psychological entitle-ment. Finally, we found that as the relational history with the same favor giver deepened, low-power (vs. high-power) favor recipients adjusted their level of gratitude expression downward, but only when communicating with similarly low-power favor givers. When interacting with higher-power favor givers, lower-power favor recipients expressed a con-sistently high level of gratitude, regardless of relational his-tory length.

Theoretical Contributions and Practical Implications

This research contributes to the power, gratitude, and impres-sion formation literatures. First, we identify a crucial ante-cedent of expressing and feeling gratitude—relative power. Whereas past work has focused on downstream conse-quences of gratitude, we empirically document power’s role in shaping gratitude feelings and expressions. Second, we identify interpersonal orientation and psychological entitle-ment as mechanisms driving these effects, thereby drawing on and integrating insights from the social distance theory of power (Magee & Smith, 2013) and the find-remind-and-bind theory of gratitude (Algoe, 2012). Third, our findings advance the perspective that power, by its very nature, is interpersonal (e.g., see Smith & Magee, 2015). Finally, we tested a novel hypothesis related to the relational history between interaction partners by leveraging insights from error management theory (Haselton & Buss, 2000). Our findings suggest that the asymmetrical costs associated with committing Type I and Type II errors may lead lower-power

favor recipients to express an especially large amount of gratitude to higher-power favor givers as a way to satisfy the punishment avoidance motive.

Our findings also have important practical implications. Expressing gratitude is a virtually costless act that has many positive effects (Bartlett & DeSteno, 2006; Cheng et al., 2015; Grant & Gino, 2010; Kaplan et al., 2014; Waters, 2012). Thus, understanding the factors that affect gratitude expression is crucially important to unleash gratitude’s potential. Our findings suggest that gratitude interventions may be most usefully implemented among powerful indi-viduals. For example, employee appreciation initiatives may lead organizational leaders to actively reflect on the ways in which their employees have benefited them, which may decrease their entitlement and increase their relational con-cern, potentially resulting in more gratitude expression.

Furthermore, leadership development programs and orga-nizational onboarding procedures should be used to educate employees about the benefits of feeling and expressing grati-tude. Given that power has been linked to self-serving behav-ior (Williams, 2014), leaders may be more receptive to gratitude interventions that are tailored to reveal the benefits leaders will receive from feeling and expressing gratitude.

Limitations

This work also has several limitations. First, we are unable to rule out alternative explanations for the findings reported in Studies 1 and 4. It is possible that low-power researchers (e.g., grad students) and Wikipedia editors (i.e., non-admin editors) may have simply received more help from others than high-power researchers (e.g., full professors) and Wikipedia editors (i.e., admins), which could explain their increased gratitude expression relative to higher-power oth-ers. Alternatively, low-power researchers and Wikipedia edi-tors may have expressed more gratitude to specific others as a way of signaling their high-status network ties to a broader audience. We address this limitation in Studies 2 and 3 by holding constant the magnitude of the favor received.

Second, Study 3 relies on a single item measure of grati-tude expression. We chose the single item measure because we believe the pre-set response options that we included accurately reflect the types of comments that individuals are likely to use. In addition, participants may have found it dif-ficult to report how much gratitude they wanted to express using a multi-item scale. We also note that Robins et al. (2001) found that a single-item measure can have high con-vergent validity with the full-item scale. Although we find consistent evidence across our set of studies, we acknowl-edge that a more involving experiment relying on a behav-ioral measure of gratitude (such as the gratitude measure used in Study 2) could shed further light on the phenomenon of interest.

Anicich et al. 17

Future Directions

Researchers should seek to understand when higher-power favor recipients are unaware of their reduced gratitude expres-sion versus when higher-power favor recipients strategically withhold gratitude. Both situations result in reduced gratitude expression, but for very different reasons. Those who are unaware of their reduced gratitude expression may be upset to learn that they have not been meeting the social needs of oth-ers, whereas those who withhold gratitude strategically may be acutely aware of the harm caused by their reduced grati-tude expression, but embrace that outcome as a means to an end. In addition, scholars should continue to study the percep-tion that higher-power individuals have of others’ generous acts and how these perceptions may affect the relationship between the two parties. Some scholars have made initial strides in this area (Inesi et al., 2012; Kunstman et al., 2018), but additional research is warranted. Finally, follow-up work could test our hypotheses in a cross-cultural context. For example, higher-power favor recipients in collectivist cul-tures compared to individualistic cultures may be more cogni-zant of others’ needs in general and the value of expressing gratitude in particular, thus leading them to express as much or even more gratitude than lower-power favor recipients.

Conclusion

Taken together, our findings identify a crucial predictor of feeling and expressing gratitude—relative power—and high-light when and why the relationship between relative power and gratitude emerges. As psychologists and practitioners alike continue to seek ways to understand and improve com-munication patterns, they would be wise to emphasize and study gratitude and the forces that affect its emergence and expression.

Acknowledgments

We are extremely grateful for the constructive feedback that the associate editor and two anonymous reviewers provided to us throughout the peer review process. In addition, we thank Adam Galinsky for providing incredibly insightful feedback on multiple versions of this paper.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, author-ship, and/or publication of this article.

ORCID iDs

Eric M. Anicich https://orcid.org/0000-0002-4043-647XShi Liu https://orcid.org/0000-0002-1941-1198

Notes

1. Data, code, and materials for all studies are available on the Open Science Framework: https://osf.io/6x9wz/?view_only=971dd6b8c2a544b49bf1acad687ce26e

2. See the SOM document for outlier detection information.3. All citation data were collected from the Business Source

Complete database on January 15, 2016.4. The data are available on Cristian Danescu-Niculescu-Mizil’s

website: http://www.cs.cornell.edu/~cristian/Echoes_of_power.html. We thank Cristian Danescu-Niculescu-Mizil and his col-leagues for making these data publicly available.

Supplemental Material

Supplemental material is available online with this article.

References

Aldridge, T. T., Audretsch, D., Desai, S., & Nadella, V. (2014). Scientist entrepreneurship across scientific fields. The Journal of Technology Transfer, 39(6), 819–835.

Algoe, S. B. (2012). Find, remind, and bind: The functions of gratitude in everyday relationships. Social and Personality Psychology Compass, 6(6), 455–469.

Algoe, S. B., Gable, S. L., & Maisel, N. C. (2010). It’s the little things: Everyday gratitude as a booster shot for romantic rela-tionships. Personal Relationships, 17(2), 217–233.

Algoe, S. B., Kurtz, L. E., & Hilaire, N. M. (2016). Putting the “you” in “thank you”: Examining other-praising behavior as the active relational ingredient in expressed gratitude. Social Psychological and Personality Science, 7(7), 658–666.

Anicich, E. M. (2016). The psychological experience of middle-power in social hierarchies: A theoretical and empirical investigation [Unpublished doctoral dissertation]. Columbia University.

Anicich, E. M., & Hirsh, J. B. (2017a). The psychology of middle power: Vertical code- switching, role conflict, and behav-ioral inhibition. Academy of Management Review, 42(4), 659–682.

Anicich, E. M., & Hirsh, J. B. (2017b, March). Why being a middle manager is so exhausting. HBR.org digital article. https://hbr.org/2017/03/why-being-a-middle-manager-is-so-exhausting

Anicich, E. M., Schaerer, M., Gale, J., & Foulk, T. A. (2020). A fluctuating sense of power is associated with reduced well-being. Journal of Experimental Social Psychology, 92, Article 104057.

Bartlett, M. Y., & DeSteno, D. (2006). Gratitude and prosocial behavior: Helping when it costs you. Psychological Science, 17(4), 319–325.

Blader, S. L., & Chen, Y. R. (2012). Differentiating the effects of status and power: A justice perspective. Journal of Personality and Social Psychology, 102(5), 994–1014.

Brown, Z. C., Anicich, E. M., & Galinsky, A. D. (2020). Compensatory conspicuous communication: Low status increases jargon use. Organizational Behavior and Human Decision Processes, 161, 274–290.

Campbell, W. K., Bonacci, A. M., Shelton, J., Exline, J. J., & Bushman, B. J. (2004). Psychological entitlement: Interpersonal

18 Personality and Social Psychology Bulletin 00(0)

consequences and validation of a self-report measure. Journal of Personality Assessment, 83(1), 29–45.

Case, C. R., Conlon, K. E., & Maner, J. K. (2015). Affiliation-seeking among the powerless: Lacking power increases social affiliative motivation. European Journal of Social Psychology, 45(3), 378–385.

Cheng, S. T., Tsui, P. K., & Lam, J. H. (2015). Improving mental health in health care practitioners: Randomized controlled trial of a gratitude intervention. Journal of Consulting and Clinical Psychology, 83(1), 177–186.

Cheung, F. M., Leung, K., Zhang, J. X., Sun, H. F., Gan, Y. Q., Song, W. Z., & Xie, D. (2001). Indigenous Chinese personal-ity constructs—Is the Five-Factor model complete? Journal of Cross-Cultural Psychology, 32, 407–433.

Cho, Y., & Fast, N. J. (2012). Power, defensive denigration, and the assuaging effect of gratitude expression. Journal of Experimental Social Psychology, 48(3), 778–782.

Côté, S., Stellar, J. E., Willer, R., Forbes, R. C., Martin, S. R., & Bianchi, E. C. (2020). The psychology of entrenched privi-lege: High socioeconomic status individuals from affluent backgrounds are uniquely high in entitlement. Personality and Social Psychology Bulletin, 47, 70–88.

Danescu-Niculescu-Mizil, C., Lee Bo Pang, L., & Kleinberg, J. (2012). Echoes of power: Language effects and power differ-ences in social interaction. In Proceedings of WWW 2012: 21st International Conference on World Wide Web (pp. 699–708). ACM.

De Cremer, D., & Van Dijk, E. (2005). When and why leaders put themselves first: Leader behaviour in resource allocations as a function of feeling entitled. European Journal of Social Psychology, 35(4), 553–563.

Emerson, R. M. (1962). Power-dependence relations. American Sociological Review, 27, 31–41.

Emmons, R. A., & McCullough, M. E. (2003). Counting blessings versus burdens: An experimental investigation of gratitude and subjective well-being in daily life. Journal of Personality and Social Psychology, 84(2), 377–389.

Fiske, S. T. (1993). Controlling other people. The impact of power on stereotyping. American Psychologist, 48(6), 621–628.

Fiske, S. T., & Berdahl, J. (2007). Social power. In A. W. Kruglanski & E. T. Higgins (Eds.), Social psychology: Handbook of basic principles (pp. 678–692). Guilford Press.

Fiske, S. T., & Neuberg, S. L. (1990). A continuum of impression formation, from category- based to individuating processes: Influences of information and motivation on attention and interpretation. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 23, pp. 1–74). Academic Press.

Galinsky, A. D., Magee, J. C., Inesi, M. E., & Gruenfeld, D. H. (2006). Power and perspectives not taken. Psychological Science, 17(12), 1068–1074.

Galinsky, A. D., Rucker, D. D., & Magee, J. C. (2015). Power: Past findings, present considerations, and future direc-tions. In M. Mikulincer, P. R. Shaver, J. A. Simpson, & J. F. Dovidio (Eds.), APA handbooks in psychology. APA handbook of personality and social psychology, Vol. 3. Interpersonal relations (pp. 421–460). American Psychological Association.

Gan, M., Heller, D., & Chen, S. (2018). The power in being yourself: Feeling authentic enhances the sense of power. Personality and Social Psychology Bulletin, 44(10), 1460–1472.

Goodwin, S. A., Gubin, A., Fiske, S. T., & Yzerbyt, V. Y. (2000). Power can bias impression processes: Stereotyping subordi-nates by default and by design. Group Processes & Intergroup Relations, 3(3), 227–256.

Grant, A. M., & Gino, F. (2010). A little thanks goes a long way: Explaining why gratitude expressions motivate prosocial behavior. Journal of Personality and Social Psychology, 98(6), 946–955.

Haselton, M. G., & Buss, D. M. (2000). Error management the-ory: A new perspective on biases in cross-sex mind reading. Journal of Personality and Social Psychology, 78(1), 81–91.

Hashimoto, H., & Yamagishi, T. (2013). Two faces of interde-pendence: Harmony seeking and rejection avoidance. Asian Journal of Social Psychology, 16(2), 142–151.

Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis. Guilford Press.

Huang, K., Yeomans, M., Brooks, A. W., Minson, J., & Gino, F. (2017). It doesn’t hurt to ask: Question-asking increases lik-ing. Journal of Personality and Social Psychology, 113(3), 430–452.

Inesi, M., Gruenfeld, D. H., & Galinsky, A. D. (2012). How power corrupts relationships: Cynical attributions for others’ gener-ous acts. Journal of Experimental Social Psychology, 48(4), 795–803.

Kaplan, S., Bradley-Geist, J. C., Ahmad, A., Anderson, A., Hargrove, A., & Lindsay, A. (2014). A test of two positive psychology interventions to increase employee well-being. Journal of Business and Psychology, 29(3), 367–380.

Kashdan, T. B., Uswatte, G., & Julian, T. (2006). Gratitude and hedonic and eudaimonic well- being in Vietnam war veterans. Behaviour Research and Therapy, 44(2), 177–199.

Kelley, H. H., Berscheid, E., Christensen, A., Harvey, J. H., Huston, T. L., Levinger, G., & Peterson, D. R. (1983). Analyzing close relationships. In H. H. Kelley, E. Berscheid, A. Christensen, J. H. Harvey, T. L. Huston, G. Levinger, & D. R. Peterson (Eds.), Close relationships (pp. 20–67). Freeman.

Kelley, H. H., & Thibaut, J. W. (1978). Interpersonal relations: A theory of interdependence. Wiley.

Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power, approach, and inhibition. Psychological Review, 110(2), 265–284.

Kunstman, J. W., Fitzpatrick, C. B., & Smith, P. K. (2018). Poisoned praise: Discounted praise backfires and undermines subordinate impressions in the minds of the powerful. Social Psychological and Personality Science, 9(4), 470–480.

Lammers, J., Galinsky, A. D., Gordijn, E. H., & Otten, S. (2012). Power increases social distance. Social Psychological and Personality Science, 3(3), 282–290.

Lee, F., & Tiedens, L. Z. (2001). Is it lonely at the top? The inde-pendence and interdependence of power holders. Research in Organizational Behavior, 23, 43–91.

Lee, H. W., Bradburn, J., Johnson, R. E., Lin, S. H. J., & Chang, C. H. D. (2018). The benefits of receiving gratitude for helpers: A daily investigation of proactive and reactive helping at work. Journal of Applied Psychology, 104(2), 197–213.

Libaers, D. (2014). Foreign-born academic scientists and their interactions with industry: Implications for university tech-nology commercialization and corporate innovation manage-ment. Journal of Product Innovation Management, 31(2), 346–360.

Anicich et al. 19

MacKenzie, M. J., & Baumeister, R. F. (2019). Motivated gratitude and the need to belong: Social exclusion increases gratitude for people low in trait entitlement. Motivation and Emotion, 43(3), 412–433.

Magee, J. C., & Galinsky, A. D. (2008). Social hierarchy: The self-reinforcing nature of power and status. The Academy of Management Annals, 2(1), 351–398.

Magee, J. C., Milliken, F. J., & Lurie, A. R. (2010). Power dif-ferences in the construal of a crisis: The immediate aftermath of September 11, 2001. Personality and Social Psychology Bulletin, 36(3), 354–370.

Magee, J. C., & Smith, P. K. (2013). The social distance theory of power. Personality and Social Psychology Review, 17(2), 158–186.

Miles, G. J., Smyrnios, K. X., Jackson, M., & Francis, A. J. (2019). Reward-punishment sensitivity bias predicts narcissism sub-types: Implications for the etiology of narcissistic personalities. Personality and Individual Differences, 141, 143–151.

Mooijman, M., Kouchaki, M., Beall, E., & Graham, J. (2020). Power decreases the moral condemnation of disgust-inducing transgressions. Organizational Behavior and Human Decision Processes, 161, 79–92.

Neria, A. L., Vizcaino, M., & Jones, D. N. (2016). Approach/avoidance tendencies in dark personalities. Personality and Individual Differences, 101, 264–269.

Pelser, J., de Ruyter, K., Wetzels, M., Grewal, D., Cox, D., & van Beuningen, J. (2015). B2B Channel Partner Programs: Disentangling Indebtedness from Gratitude. Journal of Retailing, 91(4), 660–678.

Pennebaker, J. W., Booth, R. J., Boyd, R. L., & Francis, M. E. (2015). Linguistic inquiry and word count: LIWC 2015. Pennebaker Conglomerates.

Piff, P. K. (2014). Wealth and the inflated self: Class, entitlement, and narcissism. Personality and Social Psychology Bulletin, 40(1), 34–43.

Raskin, R. N., & Hall, C. S. (1979). A narcissistic personality inventory. Psychological Reports, 45(2), Article 590.

Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global self-esteem: Construct validation of a single-item measure and the Rosenberg Self-Esteem Scale. Personality and Social Psychology Bulletin, 27(2), 151–161.

Rucker, D. D., & Galinsky, A. D. (2016). The agentic-communal model of power: Implications for consumer behavior. Current Opinion in Psychology, 10, 1–5.

Rucker, D. D., Galinsky, A. D., & Magee, J. C. (2018). The agen-tic–communal model of advantage and disadvantage: How inequality produces similarities in the psychology of power, social class, gender, and race. In J. Olson (Ed.), Advances in experimental social psychology (Vol. 58, pp. 71–125). Academic Press.

Sawaoka, T., Hughes, B. L., & Ambady, N. (2015). Power height-ens sensitivity to unfairness against the self. Personality and Social Psychology Bulletin, 41(8), 1023–1035.

Schmeichel, B. J. (2007). Attention control, memory updating, and emotion regulation temporarily reduce the capacity for execu-tive control. Journal of Experimental Psychology: General, 136(2), 241–255.

Shane, S., & Stuart, T. (2002). Organizational endowments and the performance of university start-ups. Management Science, 48(1), 154–170.

Singelis, T. M. (1994). The measurement of independent and inter-dependent self-construals. Personality and Social Psychology Bulletin, 20, 580–591.

Smith, P. K., & Magee, J. C. (2015). The interpersonal nature of power and status. Current Opinion in Behavioral Sciences, 3, 152–156.

Smith, P. K., & Trope, Y. (2006). You focus on the forest when you’re in charge of the trees: Power priming and abstract information processing. Journal of Personality and Social Psychology, 90(4), 578–596.

Spencer, S. J., Zanna, M. P., & Fong, G. T. (2005). Establishing a causal chain: Why experiments are often more effective than mediational analyses in examining psychological processes. Journal of Personality and Social Psychology, 89(6), 845–851.

Swap, W. C., & Rubin, J. Z. (1983). Measurement of interpersonal orientation. Journal of Personality and Social Psychology, 44(1), 208–219.

Twenge, J. M., & Campbell, W. K. (2009). The narcissism epi-demic: Living in the age of entitlement. Simon and Schuster.

Van Kleef, G. A., Oveis, C., van der Löwe, I., LuoKogan, A., Goetz, J., & Keltner, D. (2008). Power, distress, and compassion: Turning a blind eye to the suffering of others. Psychological Science, 19(12), 1315–1322.

Waters, L. (2012). Predicting job satisfaction: Contributions of indi-vidual gratitude and institutionalized gratitude. Psychology, 3(12), 1174–1176.

Waytz, A., Chou, E. Y., Magee, J. C., & Galinsky, A. D. (2015). Not so lonely at the top: The relationship between power and loneliness. Organizational Behavior and Human Decision Processes, 130, 69–78.

Williams, M. J. (2014). Serving the self from the seat of power goals and threats predict leaders’ self-interested behavior. Journal of Management, 40(5), 1365–1395.

Yang, H., Van de Vliert, E., Shi, K., & Huang, X. (2008). Whose side are you on? Relational orientations and their impacts on side-taking among Dutch and Chinese employees. Journal of Occupational and Organizational Psychology, 81(4), 713–731.

Zhao, X., Lynch, J. G., Jr., & Chen, Q. (2010). Reconsidering Baron and Kenny: Myths and truths about mediation analysis. Journal of Consumer Research, 37(2), 197–206.